Evaluating the effects of non tariff barriers; The economic analysis of protection in WTO disputes By Damien J. Neven University of Lausanne and CEPR April 2000 First draft – preliminary – do not quote I would like to thank Henrik Horm, Petros C. Mavroidis and Thomas von Ungern-Sternberg for useful discussions and comments on a previous draft of the paper. I am also thankful to Petros C. Mavroidis for guiding me through the WTO panels and Appellate body decisions.
Transcript
Evaluating the effects of non tariff barriers;
The economic analysis of protection in WTO disputes
By Damien J. Neven
University of Lausanne and CEPR
April 2000
First draft – preliminary – do not quote
I would like to thank Henrik Horm, Petros C. Mavroidis and Thomas von Ungern-Sternberg for useful
discussions and comments on a previous draft of the paper. I am also thankful to Petros C. Mavroidis
for guiding me through the WTO panels and Appellate body decisions.
2
Summary
This paper considers the economic analysis of the protection associated with nontariff barriers in the context of disputes under Art III of the GATT. We first observethat if Art III is meant to consider the protection of domestic firms, the appropriateconcept of protection has hardly been discussed in the case law. The circumstanceswhere protection should be measured are not defined very precisely. We also observethat panels have systematically shied away from any explicit benchmark on what couldconstitute an acceptable degree of protection and often failed to recognise thatprotection is determined jointly by the importance of the barrier and the degree ofsubstitution between domestic and foreign products.
Next, we consider a stylised model of trade where we experiment with variousmeasures of protection and investigate the interplay between the degree of productdifferentiation and the degree of rivalry in the domestic market in determining theprotection granted to domestic firms. The analysis confirms that the ability to raiseprice is a robust measure of protection and protection falls (for a given barrier) with thedegree of rivalry and the degree of substitution between products.
Finally, we suggest a method to evaluate protection in trade disputes which isinspired by the definition of the relevant market in Antitrust. We suggest that a set ofproducts should first be found in which a hypothetical monopolist controlling thedomestic supply would be able to raise price by x %. If this market cannot be found, orif it is found that competition in this market is intense, the application should bedismissed. Only if competition is weak should the investigation attempt to measure theimportance of the disputed barrier.
3
1. Introduction
Most of the economic literature dealing with non tariff barriers (like standards
and technical barriers) in international trade analyses the effects of these barriers ex ante
on the equilibrium pattern of output and welfare across countries (see Baldwin and
Venables (1999) for a survey). Over time, the multilateral system, as well as a number
of regional trading arrangements, have also established legal norms with respect to non
tariff barriers to trade. In principle, the consequences of the liberalisation associated
with the observance of these norms can also be evaluated ex post. Despite obvious
methodological difficulties1, a few studies have indeed attempted to undertake such
exercise, notably in the context of the European internal market.
This paper considers the operation of legal norm itself rather than its aggregate
consequences. As expected with any legal framework, norms give rise to disputes and
litigation. Whether a particular technical standard or non tariff barrier is lawful is
presumably not a matter that can be assessed purely on the basis of legal reasoning. To
the extent that a legal norm is not solely based on forms and relies on an assessment of
the effects of any particular measure, economic analysis will be instrumental in its
implementation. Hence, like antitrust, trade is presumably an area where legal
reasoning and economic analysis will interact.
The first part of the paper considers the legal framework towards non tariff
barriers in the multilateral system. We focus on Art III of the GATT which establishes
the principle of national treatment such that member countries are committed not to
introduce internal measures which discriminate in favour of domestic products. This
provision is the main legal norm against which non tariff barriers have to be assessed.
We review both the legal norm itself and some of the most influential case law. We
first observe that if the language of the Art III would seem to call for an evaluation of
1 As it is necessary to establish a reasonable counterfactual (what would have happened without the legalnorm) against which actual observations can be assessed.
4
the effects of disputed measures, the case law has sometimes consciously not relied on
an assessment of effects and has thus neglected or avoided an economic evaluation. For
the majority of cases however the usefulness of an economic approach is recognised.
For those cases, the economic analysis which is proposed is still remarkably
unstructured. This is partly because Art III does not clarify how effects should be
assessed. Art III expresses concern about protection of domestic firms but it is not
clear how protection should be measured and in particular whether the effects of a
disputed instrument should be measured in terms of price, trade flows or rents for the
domestic firms.
Some structure in the investigation still arises from the need to find a market as a
reference. Indeed, the investigation has to decide which domestic products are "like" or
"directly competitive and subsitutable" with the imported item which is allegedly
discriminated against.
The case law that we review still does not follow a systematic approach in the
delineation of this market and the approach that is followed often suffers from important
shortcomings. This case law also reveals that panels have avoided making reference to
an benchmark for the evaluation of protection. Yet, protection is matter of degree and
one cannot escape from the definition of what should be considered as an acceptable
degree of protection.
Overall, this state of affairs is in stark contrast with antitrust investigations
where objectives are reasonably clear and where experienced agencies follow a
structured set of principles in their analysis. In particular, antitrust investigations
often proceed by first delineating the relevant market according to clear principles and
few would dispute that a rigorous market definition greatly contributes to the clarity and
ultimate quality of antitrust decisions.
The remainder of the paper then seeks to develop a framework to evaluate the
effects of disputed measures. Section 3 develops a simple model of trade with
differentiated products and considers alternative variables which can be used to measure
5
the effects of barriers which raise the (marginal) cost of foreign firms. We conclude
that effects on domestic (or foreign) prices are more consistent and reliable than effects
on the value of trade. We also observe that the extent of protection, as measured by the
ability to raise price for the domestic firms (for a given barrier), is determined by the
degree of substitution between domestic and foreign products and by the degree of
competition between them. A strong similarity between domestic and foreign products
will thus not be a sufficient condition for the former to obtain protection from a
differential treatment. When competition is intense, domestic firms will simply not
benefits from a differential treatment imposed on foreign firms.
In light of these findings, section 4 outlines a simple procedure to evaluate the
effects of disputed measures. We propose a two step procedure. In the first step, a
set of products will be found such that a hypothetical monopolist controlling the
domestic market would be able to raise price by x % above the competitive level. If it
is found that no such market exists (because for instance demand substitution is large or
foreign competition is strong despite the barrier) , the case should be dismissed. If such
a market can be found, the second step of the evaluation would be to consider
competition in relevant market. If it is found that competition is intense, the case
should again be dismissed as no significant effect will be expected. If competition is
not intense (for instance when concentration and barriers to entry are high), the
investigation should then consider to what extent domestic prices would fall if the
additional cost imposed on foreign firms by the disputed barriers would be removed.
2. Economic analysis in Art III of GATT
Art III of the GATT is concerned with situations where domestic regulations
including taxes are applied in such a way as to discriminate in favour of domestic firms.
This is a general provision which apply in a wide range of circumstances. We will first
briefly discuss (section 2.1) its scope and in particular whether technical standards or
health related standards could be assessed against the norm of Art III. We will then
(section 2.2.) examine whether Art III contains a test in terms of the effects of disputed
barriers and how the case law has defined the set of products that are "like" or "directly
6
competitive and substitutable" with the foreign product subject to the disputed
measures. Some conclusions are offered in section 2.3.
2.1. Some remarks on the scope of Art III
Art III expresses a clear norm which respect to discrimination in favour of domestic
firms which in principle covers all internal measures. Discrimination can still be
justified by appealing to the provisions of Art XX. The TBT and SPS agreements
also introduce additional commitments with respect to technical standards and sanitary
(or Phyto-sanitary) standards. In essence, the TBT agreement commits the WTO
members to use international standards when possible and not to introduce national
standards which favour domestic production. The agreement on Sanitary and Phyto-
Sanitary products establishes the principles that WTO members need to produce
scientific evidence when they appeal to the protection of health in order to justify a
trade restrictive measure. These two agreements thus provide specific arguments (in
lieu of those contained in Art XX) that can justify discriminatory measures. Technical
barriers covered by the TBT can thus be challenged in terms of their conformity with
respectively the TBT and SPS agreements. These actions will then typically question
the validity of the justification behind the trade restrictive measures (as in the Hormones
and Australian Salmon cases in the SPS context for instance). However, technical
barriers can be (and have) also challenged under Art III. Even though this is so far
untested, the same could in principle happen with health related measures.
2.2. “Like”, “directly competitive and substitutable products” and protection
As indicated above, Art III.1 is concerned with internal measures which give a
different treatment to foreign and national products. Art III.1 informs the domestic
regulation with respect to both the treatment of charge elements (III.2) and non charge
elements (III.4). Its concern is expressed in the following terms : it suggests that taxes
or internal measures “should not be applied to imported or domestic products so as to
afford protection to domestic production” (see appendix for the full text). Hence, the
article emphasises the protection that is granted to domestic firms and makes no
7
reference to the cost imposed on foreign firms. Accordingly, an investigation of a
possible violation of Art III should normally focus on domestic firms. What is meant
by protection and how it should be measured is however not clarified. It is not clear
either whether the form (the object) of a measure is a sufficient ground to conclude that
it is unlawful.
With respect to Art III.2 (charge elements), some further insight is provided by
the panel decision on Japanese Alcoholic beverages (confirmed by the Appellate body
decision on this point). According to this decision, Art III is not concerned with intent
and what matters is not whether member governments designed their domestic
regulation in order to favour domestic firms but only whether domestic regulation does
actually grant protection.
Art III still provides some structure for the investigation of domestic protection.
It draws a distinction between domestic products which can be qualified as “like”
products relative to those imported (first sentence of para 2) and domestic products that
are “directly competitive and substitutable ” for imported ones (second sentence of para
2). These two sets of products are subject to different legal standards2.
Some insight is provided by the panel decision in Superfund. In this case, the
US government had provided evidence that the disputed measure had no significant
effect on trade. The panel did not question the evidence but dismissed the argument
altogether. The panel concluded, interpreting literally article III.2, that with respect to
“like” products, even the slightest deviation is enough to establish discrimination
against foreign products. Hence, it would seem that with respect to charge elements
and with respect to “like” products, there must be absolute identity with respect to the
treatment reserved for domestic and foreign product. The panel first argued that
evidence on trade effects was not relevant because one could not determine the
appropriate counterfactual against which changes in trade flows should be measured. In
case of deviation (heavier tax on foreign products), there is no need for the panel to
check the effects of the deviation. The panel noted that several solutions could be
2 See Mavroidis (2000) or Bronckers and McNelis (2000) for a discussion of this issue.
8
adopted to bring the disputed tax in conformity with Art III, including an increase in the
tax imposed on US products to bring it in line with the tax imposed on foreign products
and a reduction in the tax imposed on foreign products to bring it into line with the US
internal level. According to the panel, “each of these solutions would have different
trade results and it is therefore logically not possible to determine the difference in trade
impact between the present tax and one consistent with Art III.2”. This argument is
odd for at least two reasons. First, it is very likely that the net trade impact associated
with the differential treatment that one would measure in the alternative scenarios
mentioned by the panel would be very similar. Second, and more fundamentally, it
seems odd to dismiss evidence simply because it is subject to some strictly positive (but
arbitrarily small) measurement error.
The Panel concluded that Art III “cannot be interpreted to protect expectations
on exports volume; it protects expectations on the competitive relationship between
imported and domestic products”. This statement is hard to understand. Indeed, one
wonders how the competitive relationship which is supposedly protected by Art III will
be assessed if trade effects are not relevant because they cannot be measured precisely.
Surely, whether the competitive relationship is affected cannot be inferred “logically”
(unless of course the competitive relationship is defined very formally - in terms of the
difference in regulation). It will have to be evaluated in terms of effects and those will
presumably not be evaluated with more precision than trade effects.
Hence, it appears when with respect to like products and charge elements, Art
III does not require an evaluation of effects. The legal standard for the treatment of
imported and like products (for charge element) is thus particularly strict as it admits no
difference of treatment across products. This standard has also been affirmed without
ambiguity by the Appellate body in the case of Japanese alcoholic beverages ; according
to the Appellate body, “even the smallest amount of excess is too much… The
prohibition under article III2, first sentence, is not conditional on a trade effects test nor
is it qualified by a de minimis standard”.
9
This begs the question of how “like” products are determined and in particular
how “like” products are distinguished from “directly competitive and substitutable
ones”.
As clarified by various panel decisions, “like” products form a subset of
“directly competitive and substitutable” ones. Imported and like domestic products are
supposed to share most physical characteristics and should normally be subject to the
same HS tariff heading.
With respect to “competitive and substitutable products”, the legal standard is
less clear-cut. First, it appears that differences in the treatment of domestic and foreign
products is subject to a de minimis standard. Hence, in order to find a prohibition, it
will be necessary to show that the treatment of domestic and foreign products is not
similar. Second, since the second sentence of art III.2 makes explicit reference to the
first paragraph of the same article3, it appears that in order to find a prohibition, it will
also be necessary to show that differences in the treatment of foreign and domestic
products do effectively grant protection to the latter.
The distinction between “like” and “directly competitive and substitutable”
products and the very strict standard applied to the former relative to the latter seems to
be associated with the presumption that if two products are very similar, any difference
in the treatment applied to them will translate into a significant protection for the
product which benefit from more favourable conditions. In such case, it can simply
be presumed that the protective effect exists and is significant. By contrast, when
products are more differentiated, the significance of the protective effect has to be
established.
The practical significance of the distinction between different legal standards
applied to “like” and “directly competitive and substitutable” products could be
questioned. First, the distinction between the two types of products is in practice not
easily drawn. According to various panel decisions, the distinction has to be made on
3 See appendix.
10
case by case basis and there is not a single criterion that should be used in all cases.
As indicated above, there is an apparent consensus that “like” products should share
most physical characteristics. With respect to the assessment of whether products are
directly competitive and substitutable, there is repeated suggestion in panels that it
should be determined in the market place and that similarity in end-use (as measured
possibly by the elasticity of substitution across products or the cross-price elasticity of
demand) is an important factor to take into account. Beyond this, little systematic
guidance is offered by the panel decisions. The (unusually lyrical) comment by the
Appelate body on Japanese alcoholic beverages (p 23) that the “concept of likeness is
one that evokes the image of an accordion… (which) stretches and squeezes in different
places as different provisions of the WTO agreements are applied” also offers little
comfort that the distinction between the two types of products can be established on a
sound and systematic basis.
Second, the distinction between the legal standards applied to the two types of
products seems (at least so far) largely immaterial. The case of Japanese alcoholic
beverages is particularly telling in this respect : the panel decided (p120) that “for it to
conclude that dissimilar taxation afforded protection, it would be sufficient for it to find
that the dissimilarity is not de minimis”. In other words, if the difference of treatment is
large enough the protective effect can simply be presumed (as in the case of “like”
products). Hence, the only difference between the standard applied to “like” and
“directly competitive and substitutable” products relates to the importance of the
dissimilarity in treatment. Admittedly, the Appellate body disagreed with this
approach and insisted that the issue of whether protection was granted should be
handled separately from the issue of whether the treatment was dissimilar.
Unfortunately, however, the Appellate body only provide vague guidance on how to
assess protection. It merely referred to the fact that beyond the importance of the
dissimilarity in treatment, there will be “other factors to be considered”. What these
factors should be has unfortunately not been clarified by the Appellate body.
11
2.3. Some conclusions
A few conclusions can be drawn from this brief overview4. First, it is striking
while no evaluation of effects is required for “like” products, some evaluation is
required for “directly competitive and substitutable products”. At the same time, the
distinction between the two is far from clear and by itself such distinction would seem
to presuppose the existence of an effects test. The mere fact that Art III introduces a
distinction between two levels of substitution (on the one hand, “like products” and on
the other hand, “directly competitive and substitutable products”) would seem to imply
that it contains an effects test. It is also striking that how protection should be
measured has a not been seriously discussed so far.
One should emphasise that without a requirement to consider the effects of
disputed measures, Art III would become a very formal instrument. A contrario, an
emphasis on effects would make Art III potentially very powerful and in this event, a
rigorous and systematic approach to evaluate whether domestic firms are effectively
protected by the implementation of a domestic regulation will be required. Of course,
some panels may have shielded away from considering effects precisely because they
did not feel well equipped to evaluate them.
Second, the overall approach being followed by the Panels is odd, as revealed
by our discussion of the distinction between “like” and “directly competitive and
substitutable products”. It is striking that panels avoid quantitative benchmarks. Even
though the extent to which products are substitute for one another is obviously a matter
of degree, there is no benchmark to decide when this substitution is low enough to
consider that products are not directly competitive. By the same token, panels have not
considered any benchmark to evaluate the dissimilarity in products. Panels have also
tried to identify the products that are “like” or “directly competitive and substitutable”
separately from their evaluation of the importance of the barriers which are raised
4 Our discussion has focused exclusively on charge elements (Art III.2). The case law with respect tonon charge element is very scarce. According to Mavroidis (2000), an effects test should be read in theprovision (Art III.4) dealing with non charges.
12
towards foreign firms (the importance of the dissimilarity in treatment). Still, one
should recognise that protection is indeed a matter of degree and not a matter of
principles. In addition, the extent to which domestic firms will be protected by a non
tariff barrier to trade will be determined (among other things) by the degree of
substitution between domestic and foreign products and by the importance of the barrier
which is raised. In other words, if Art III expresses a standard with respect to the
degree of protection, one cannot escape from the expression of some benchmark for
what is considered an admissible degree of protection. In addition, the evaluation of
actual protection against this benchmark should consider jointly the importance of the
barrier and the degree of substitution between domestic and foreign products.
In the next section we further clarify the link between the degree of protection
that is granted to domestic firms as a function of the importance of the barrier imposed
on foreign firms, the substitution between foreign and domestic products and the degree
of competition in the domestic market. We also experiment with several measures of
protection.
3. Determinants of protection – A stylised model
The objective of this section is to compare alternative measures of protection
and to analyse the interplay between the different factors which determine the degree of
protection enjoyed by domestic firms. We build a stylised model of competition to
investigate this issue. In line with most of literature, we model non tariff barriers as an
increase in the marginal cost of foreign firms. This is certainly appropriate for some
barriers like national quality standards or national product specification which require a
change in the production methods of foreign firms. It is clear however that some
technical barriers (like national certification procedures) do involve a fixed cost and not
necessarily affect marginal cost.
To the extent that the measures concerned do indeed raise the cost of foreign
rivals, the evaluation of protective effects thus requires to identify the extent to which
13
domestic firms are protected by a change in the marginal cost of a subset of their
competitors.
This question is actually closely related to the analysis of anti-competitive
strategies aiming at raising rivals' cost which has been considered in the IO and
Antitrust literature (see Salop and Scheffman, (1987, Scheffman (1992) or
Krattenmaker and Salop (1986)). The issue there is whether and to what extent a group
of firms could increase rents and profits by raising the cost of their rivals. For instance,
Salop and Scheffman (1987) consider a Cournot model of competition and analyse the
extent to which a change in cost which is asymmetric (i.e. which affects some firms
more than others) can actually benefits the firms which are less affected by the cost
increase. One difference between the question raised here and that analysed in this
literature is of course that incumbents firms in our framework do not bear any cost in
order to raise barriers for their rivals. The cost increase of rivals is handed to hem free
of charge by their domestic government.
The first question which arises is then how to measure the protection granted to
domestic firms. As discussed above, the legal framework offers little guidance in this
respect. One natural candidate would be the rents accruing to domestic firms (relative
to what they would earn in the absence of a measure raising the marginal cost of the
foreign firms). This may however not be realistic. Profits and rents are notoriously
difficult to measure (because of accounting conventions which blurr economic profits
but more fundamentally because much of the rent accruing to firms is likely to translate
into higher cost rather than higher profit). To evaluate changes in profits relative a
counterfactual which is not observed is of course even harder and this can only be
undertaken at great cost – and with a substantial degree of imprecision. Profits and
rents are thus hardly used even in antitrust investigations and it would equally difficult
to use them in trade investigations. One alternative candidate to measure protection
would the importance of the price distortion associated with protection, namely the
price that domestic firms are able to charge relative to the level that would prevail
without the increase of the marginal cost of foreign firms. The price that foreign firms
can charge relative to what they would charge in the absence of the barrier is another
14
possibility. Finally, the value of imports (relative to what would be observed without
the increase in the marginal cost of foreign firms) could be considered. As mentioned
above, this measure what used in the Superfund case.
Let us mention two advantages of price measures at the outset. First,
counterfactuals over prices are routinely used in antitrust investigations which suggests
that a test based on prices could be implemented from a practical point of view. Third,
the focus on price distortions is also well in line with the stated objective of the GATT
agreements, whose preamble emphasises an efficient allocation of resources5.
Intuitively, one would expect that the extent to which domestic firms will be
protected will be affected by the degree of differentiation between domestic and foreign
products.. Intuitively, if domestic and foreign products hardly differentiated, domestic
firms will benefit from an increase in the cost of their foreign competitors ; in those
circumstances, the cross price elasticity of demand for the domestic product with
respect to the foreign price will be high so that when the foreign firm raises its price
following its rise in cost, it will lose a substantial proportion of its customers to the
domestic firms. In turn, domestic firms will also tend to have a strong incentive to raise
their own price. The extent to which domestic firms benefit might still depend on the
type of competition taking place and in particular whether firms are selling strategic
substitutes or strategic complements.
If the demand substitution between foreign and domestic products should be an
important determinant of protection, the extent of rivalry which prevails between firms
will also matter a great deal. For instance, consider an extreme case where there is
perfect competition between firms (see figure 1). Assume that domestic firms operate
with a constant marginal cost so that the supply curve of the domestic industry (DS) is
horizontal. Foreign firms have an upward sloping supply curve (FS). As their supply
curve is shifted to the left by the discriminatory measure (to FS’), the equilibrium price
in the domestic market does not change. All what happens is that the market share of
the domestic firms increase at the expense of foreign firms and domestic firms do not 5 At least, abstracting from second best considerations.
15
earn additional rents (in perfect competition and constant marginal cost, producer
surplus is anyway zero). Admittedly, this example is a bit constricted but it illustrates
that even if substitution is perfect, protection can be absent.
Figure 1. Product differentiation and protection
In order to investigate this issue further, consider the following stylised model of
international competition6. Consider a market where products are differentiated and
where each firm is selling a different variety. There are (n-1) firms in the domestic
market and there is one foreign firm7. Domestic firms operate with a marginal cost of
c whereas the foreign firm is subject to a differential treatment such that its marginal
cost, denoted Fc is higher than c.
6 The analysis presented here can be seen as an extension of Salop and Scheffman (1987) in which weallow for product differentiation.7 The assumption that there is a single foreign firm is unimportant. Results would not change if therewere m < n foreign firms and (n-m) domestic firms.
DS
DD
FS
FS ’p
q
16
Demand for the differentiated products is given by the following system (see
Shubik and Levitan (1980) ) which allows for a fairly general pattern of symmetric
substitution across products :
where γ is a measure of the substitution across products ( 0=γ describes independent
products and the degree of substitution increases with the parameter – for ∞=γ ,
products are homogenous).
In order to trace out the effect of differential treatment on protection, we derive
the change in the price charged by domestic firm, the change in the price charged by the
foreign firm and the value of imports as a function of the marginal cost of the foreign
firm. We consider both the case of strategic substitutes and the case of strategic
complements. Let us first consider strategic complements, i.e. price competition.
The profit function of firm i is written :
Π ii i
i jj
nP cn
v Pn
P= − − + +
=∑( )1
1
γ γ
The maximisation of profit with respect to price yields the following FOC for each firm:
v Pn
P P cni j
j
n
j i− + + − − + − ==∑( ) ( )( )1 1 0
1
γ γ γ γ i=1,...,n
Since all domestic firms are symmetric and charge the same price in equilibrium, this
system of equation reduces to a couple of FOCs, one of the domestic firm and one for
the foreign firm which jointly determine the price of domestic firms P and the price
of the foreign firm PF :
++−= ∑ ji P
nPv
nq
γγ)1(1
0
17
0)1)(()1()1( =−+−−+−++−n
cPPn
Pnn
Pv Fγγγγγ
0)1)(()1()1( =−+−−+−++−n
cPPn
Pnn
Pv FFFFγγγγγ
Solving this system for the equilibrium prices *P and PF* , one obtains :
( ) ( ) ( )( )( )( ) ( )( )222
22*
322341212γγγγγ
γγγγγγγ+++++
+++++++++=nn
nncncvcnvncP FFF
( ) ( )( ) ( ) ( )( )( )( ) ( )( )222
22222*
322343234221
γγγγγγγγγγγγγγ
++++++++++++++−+−=
nnnnccvncvcn
P FF
It is then easy to check (see appendix 2) that
,0,0
,0,0
,,0,0
*2*2
*2*2
****
>>
<<
<>> −
∂γ∂
∂
∂γ∂
∂
∂∂
∂
∂∂
∂
∂
∂
∂
∂
∂
∂
∂
∂
F
D
F
F
F
D
F
F
F
F
D
F
F
F
c
P
c
P
nc
P
nc
P
c
P
c
P
c
P
c
P
Hence, when a differential treatment is applied to the foreign firm, so that its
marginal cost increases, its price increases. The price of domestic firms also increases
but by less. The increase in price is also larger, the smaller is the number of firms and
the larger is the degree of substitution across products (the lower is product
differentiation).
18
Looking at the value of imports (I), it appears that value of imports does not
always fall (as one would expect) when the cost of the foreign firm increases (see
appendix 2).
As mentioned above, in order to check the robustness of alternatives measures
of protection, we also considered the possibility of Cournot competition (i.e. strategic
substitution). The demand system given above can be inverted and the inverse demand
is then written as follows :
Using lower case letters to denote prices and quantities in the Cournot game, one
obtains the following first order conditions for respectively each of the domestic firm
and the foreign firm:
( )( )( ) 01
q2qncv F =+
++−−γ
γγ
( ) ( )( )( ) 01
n2q1nqcv FF =+
++−−−γ
γγ
Solving this system of equations for equilibrium quantities and computing the resulting
equilibrium prices, on obtains :
( ) ( )( ) ( )( )( ) ( )( )γγγ
γγγγγ+++
+++−++++=2nn2
1n2ncvcnv2np
222F*
( )( ) ( )( )( ) ( )( )γγγ
γγγγγγγ+++
++++++−+=2nn2
2nn2cvcncnv2np
22F*
F
The comparative statics of equilibrium prices with respect to the cost of the
foreign firm is then written :
[ ]∑++
−= ji qnqvp γγ)1(
10
19
( )( ) ( )( )γγγ
γγ+++
+=∂∂
2nn2n
cp
F
*
( )( ) ( )( )γγγ
γγγ++++++=
∂∂
2nn22nn2
cp 22
F
*F
We observed that, as in the case of price competition, equilibrium prices of both
the domestic and the foreign firm increase when the cost of the foreign firm increases.
The effects of the number of firms and of the degree of product differentiation are also
the same as before (see appendix 2). The smaller is the number of firms and the lower
is product differentiation, the higher is the protection enjoyed by domestic firms in
terms of price. With respect to import values, the effect is now without ambiguity. The
value of imports fall when barriers are increased.
Overall, this analysis confirms that the higher is domestic rivalry (as induced by
a large number of competitors), the lower will be the protection, as measured by the
ability to increase price, granted by a discriminatory measure against foreign firm.
Importantly, this suggests that the distinction drawn by article III of the GATT between
“like” and “directly competitive and substitutable” products is unlikely to be very
useful. The presumption that “like” products will always be protected is simply not
supported by economic principles. When competition is intense, even "like" products
will not enjoy much protection.
The analysis also confirms that price is a robust measure of protection.
Protection as measured by price is a monotonic increasing function of the cost
disadvantage suffered by the foreign firm (for both price and quantity competition). By
contrast, the value of imports appears to be less robust. With price competition, the
value of imports is not a monotonic function of the cost level of the foreign firm. For
some (possibly limited) range of parameters, the increase in price for the foreign firm
will actually more than compensate for the reduction of quantity that it sells in the
domestic market and total imports will increase. Accordingly, price appears to be a
20
better measure of protection than trade flows and in what follows we focus on the
former.
4. A proposal for the evaluation of protection
Evaluating protection in terms of price will thus require to set a benchmark
expressing an upper limit on the ability to raise price for domestic firms which is
considered acceptable. The question will then also arise of how to evaluate in practice
the extent to which a disputed measure allows (or would allow) domestic firms to raise
price.
A comparison with Antitrust proceeding is instructive at this stage. In Antitrust
proceedings, the authorities need to evaluate whether firms can effectively exercise
market power, i.e. raise price above the competitive level. This evaluation is usually
undertaken by first calibrating the market; that is, authorities proceed by first
delineating the relevant Antitrust market which is the set of product over which a
monopolist could exercise a given degree of market power. This is usually expressed as
the ability to profitably increase price by 5-10 % (see for instance the US Merger
Guideline or the EU notice on market definition). This market is defined both in terms
of product characteristics and in terms geographical scope. The main benefit from
this approach is two-fold: first, the procedure sets an explicit upper limit on the level
of market power. An ability to increase price by 5-10 % above the competitive level is
considered excessive. Second, the procedure ensures that the competitive analysis will
be meaningful. By making sure that a market share of 100 % (that of a monopolist)
would lead to the exercise of market power, the level of market power associated with
lower level of market shares and concentration can be inferred.
As discussed in the previous section, the protection enjoyed by domestic firms,
for a given barrier (cost disadvantage for the foreign firm) will be determined by the
substitution between domestic and foreign products and by degree of rivalry between
domestic firms. Hence, in order to make sure that the analysis of domestic competition
is meaningful, it may be useful, as in Antitrust proceedings, to delineate first the market
21
over which in the absence of competition, domestic firms would benefit from the
existence of the barrier; that is, it may be useful to first delineate the set of product over
which a domestic monopolist could obtain a given degree of protection. If price is
taken as the measure of protection, this would entail finding a set of products such that a
domestic monopolist controlling these products would be able to raise price above the
competitive level by some percentage. The percentage which is chosen would give an
explicit benchmark about the level of protection which is considered excessive. The
analysis should also consider that factors that are considered in antitrust proceedings; it
should consider both demand and supply substitution.
At this stage two possibilities can arise. On the one hand, it is entirely possible
that there is no set of products over which a domestic monopolist would be able to raise
price. This might happen for instance if demand substitution is important (so that
consumer will switch away and purchase products abroad if the domestic price is
increased) or if the disputed barrier is unimportant and foreign competition is intense
(so that supply substitution is significant). If no such market can be found, it implies
that even a domestic monopolist could not obtain protection and the case should be
dismissed (see figure 2).
On the other hand, if a market has been identified where a domestic monopolist
could obtain protection, one should proceed by undertaking a competitive analysis in
this market. The analysis should follow the usual approach of the evaluation of
dominance in antitrust. If it is found that competition is intense, for instance because
concentration is low and entry barriers are unimportant, there will be again a strong
presumption that domestic firms will not be able to obtain protection. In this event,
the case should again be dismissed.
If competition is weak, the investigation should proceed by analysing the extent
to which the disputed measure enables the domestic firms to raise price. This would
entail considering explicitly the increase in cost associated with the imposition of the
differential treatment on foreign firms. If it is found that the change in cost is
significant, one should conclude that the disputed measure grants a significant
22
protection to domestic firms. What is meant by a “significant” increase in cost should
also be considered in line with the benchmark which is chosen for the potential increase
in price. For instance, if an price increase of 5 % is considered at the market definition
stage, larger percentages of cost increases should be considered as significant (given
that the pass-through is less than 1, even when the domestic market structure is very
concentrated).
5. Conclusion
To sum up, this paper considers the economic analysis of the protection
associated with non tariff barriers in the context of disputes under Art III of the GATT.
We first observe that if Art III is meant to consider the protection of domestic firms, the
appropriate concept of protection has hardly been discussed in the case law. The
circumstances where protection should be measured are not defined very precisely. We
also observe that panels have systematically shied away from any explicit benchmark on
what could constitute an acceptable degree of protection and often failed to recognise
that protection is determined jointly by the importance of the barrier and the degree of
substitution between domestic and foreign products.
Next, we consider a stylised model of trade where we experiment with various
measures of protection and investigate the interplay between the degree of product
differentiation and the degree of rivalry in the domestic market in determining the
protection granted to domestic firms. The analysis confirms that the ability to raise
price is a robust measure of protection and protection falls (for a given barrier) with the
degree of rivalry and the degree of substitution between products.
Finally, we suggest a method to evaluate protection in trade disputes which is
inspired by the definition of the relevant market in Antitrust. We suggest that a set of
products should first be found in which a hypothetical monopolist controlling the
domestic supply would be able to raise price by x %. If this market cannot be found, or
if it is found that competition in this market is intense, the application should be
23
dismissed. Only if competition is weak should the investigation attempt to measure the
importance of the disputed barrier.
24
References
Baldwin, R. and T. Venables, (1999),
Bronckers, M., and N. McNelis, (2000), Rethinking the “like” product definition in
GATT: antidumping and environmental protection, Regulatory barriers and the
principle of non discrimination in world trade law, World Trade Forum Series,
Michigan University Press
Krattenmaker, T. and S. Salop, (1986), Anticompetitive exclusion: raising rival’s costs
to achieve power over price, The Yale Law Journal, vol 96, 2, 209-293
Mavroidis, P., (2000), The principle of non discrimination : an analysis at the positive
level, Regulatory barriers and the principle of non discrimination in world trade law,
World Trade Forum Series, Michigan University Press
Salop, S. and D. Scheffman, (1987), Cost-Raising Strategies, The Journal of Industrial
Economics, Vol XXXVI, 1, 19-34
Scheffman, D., (1992), The application of raising rivals’ costs theory to antitrust, The
Antitrust Bulletin, 187-206.
Shubik, M. and R. Levitan, (1980), Market structure and behaviour, Harvard University
Press
25
Appendix 1
Article III*
National Treatment on Internal Taxation and Regulation
1. The contracting parties recognize that internal taxes and other internal charges,and laws, regulations and requirements affecting the internal sale, offering for sale,purchase, transportation, distribution or use of products, and internal quantitativeregulations requiring the mixture, processing or use of products in specified amounts orproportions, should not be applied to imported or domestic products so as to affordprotection to domestic production.*
2. The products of the territory of any contracting party imported into the territory ofany other contracting party shall not be subject, directly or indirectly, to internal taxes orother internal charges of any kind in excess of those applied, directly or indirectly, to likedomestic products. Moreover, no contracting party shall otherwise apply internal taxes orother internal charges to imported or domestic products in a manner contrary to theprinciples set forth in paragraph 1.*
Ad Article III
Any internal tax or other internal charge, or any law, regulation or requirement of thekind referred to in paragraph 1 which applies to an imported product and to the likedomestic product and is collected or enforced in the case of the imported product at thetime or point of importation, is nevertheless to be regarded as an internal tax or otherinternal charge, or a law, regulation or requirement of the kind referred to in paragraph 1,and is accordingly subject to the provisions of Article III.
Paragraph 1
The application of paragraph 1 to internal taxes imposed by local governments andauthorities with the territory of a contracting party is subject to the provisions of the finalparagraph of Article XXIV. The term "reasonable measures" in the last-mentionedparagraph would not require, for example, the repeal of existing national legislationauthorizing local governments to impose internal taxes which, although technicallyinconsistent with the letter of Article III, are not in fact inconsistent with its spirit, if suchrepeal would result in a serious financial hardship for the local governments or authoritiesconcerned. With regard to taxation by local governments or authorities which isinconsistent with both the letter and spirit of Article III, the term "reasonable measures"would permit a contracting party to eliminate the inconsistent taxation gradually over atransition period, if abrupt action would create serious administrative and financialdifficulties.
Paragraph 2
26
A tax conforming to the requirements of the first sentence of paragraph 2 would beconsidered to be inconsistent with the provisions of the second sentence only in caseswhere competition was involved between, on the one hand, the taxed product and, on theother hand, a directly competitive or substitutable product which was not similarly taxed.
_______________ ¹This Protocol entered into force on 14 December 1948.
27
Appendix 2
For strategic complements, the comparative statics of the equilibrium prices are given
by :
( ) ( )( ) ( )( )222
222
F
*F
32n234n32n34n2
cP
γγγγγγγγγγ
++++++++++=
∂∂
( )( ) ( )( )222
F
*
32n234nnn
cP
γγγγγγγγ
+++++++=
∂∂
The marginal effects of the number of firms and of the degree of product differentiation
on the price of the domestic firms can then be written as :
( ) ( )( )( ) ( )( )2222
222
F
*2
32n234n
78n1n43nc
P
γγγγγγγγγ
γ +++++++++=
∂∂∂
( ) ( ) ( ) ( )( )( ) ( )( )2222
2222
F
*2
32n234n
32n42321n2nc
P
γγγγγγγγγγγγγ
++++++++++++−=
∂∂∂
And for the foreign firm :
( ) ( )( )( ) ( )( )2222
222
F
*F
2
32n234n
5128n34n34nc
P
γγγγγγγγγγ
γ ++++++++++=
∂∂∂
( ) ( ) ( )( )( ) ( )( )2222
32222
F
*F
2
32n234n
313188n32n634nc
P
γγγγγγγγγγγγγγ
+++++++++++++−=
∂∂∂
With respect to the value of imports, one obtains that
